Many piezometers/wells produce at such low rates and frequency that th
eir impact on local hydraulics is negligible, e.g. monitoring piezomet
ers or homeowner wells installed in highly permeable soils. Defining a
deterministic capture zone for such wells is often of limited utility
as the capture zone is effectively a single flow line directed upgrad
ient. In order to provide more insight into the capture zone of such w
ells, a statistically based capture zone, termed a ''percentile captur
e contour'' (PCC), is introduced and analyzed. The capture zone is def
ined by quantifying, for a given travel time, the variation of the len
gth and orientation of the flow line emanating upgradient from the wel
l. Capture zone variation herein depends on second-order stationary ra
ndom hydraulic conductivity fields and is calculated using Monte Carlo
analysis. Monte Carlo analysis yielded increases in mean travel dista
nces as the variance and/or Integral scale in log K increased, but dec
reases as the angle between the principal direction of the correlation
structure and the regional flow increased. The average travel distanc
e exceeded the travel distance estimated by a homogeneous solution usi
ng the geometric mean hydraulic conductivity. Transverse variation dep
ended both on the variance and integral scale of log K, but was insens
itive to the orientation of the principal correlation direction. The m
ean orientation of the flow path varied with the principal direction o
f the correlation structure, deviating up to 20 degrees from the orien
tation of the hydraulic gradient. These observations are consistent wi
th flow following preferential pathways and indicate that significant
uncertainty exists for source prediction of water feeding passive well
s.